Table-Us

Table-Us is an AI-powered social dining app that helps people discover authentic local food, align group preferences, and decide where to eat faster together.

Inspiration

We were inspired by a simple but frustrating experience: finding food with other people should be fun, but it often turns into a long, messy decision process. Existing food apps are great at showing ratings, reviews, and endless restaurant lists, but they do a poor job of helping people answer the real question: where should we eat together right now?

We also noticed a deeper problem. Food is one of the most direct ways people experience local culture, but most apps flatten that experience into generic rankings and SEO-optimized results. Instead of helping people discover places that match their tastes, context, and group dynamics, they push users toward the same obvious choices. We wanted to build something that felt more human, social, and culturally aware.

What it does

Table-Us helps users make dining decisions together.

The app lets people express cravings and preferences more naturally, including taste, vibe, dietary needs, price sensitivity, and social context. Instead of forcing users to scroll through endless restaurant lists, Table-Us uses AI to interpret what the group actually wants and turn that into better recommendations.

Our goal is to solve three problems at once:

Discovery overload — too many restaurant options, too little confidence
Group indecision — endless “where should we eat?” chats
Shallow personalization — apps know ratings, but not real taste profiles

With Table-Us, dining becomes less about searching and more about matching people, mood, and place.

How we built it and AI's role

We designed Table-Us as an AI-native product rather than a traditional restaurant directory.

On the frontend, we used React / Next.js with TypeScript and Tailwind CSS to build the interface and prototype the user experience quickly. We focused on flows for expressing preferences, exploring restaurant options, and making group decisions.

On the backend, we used Node.js and integrated LLM-based reasoning to interpret user intent beyond keywords. Instead of only relying on filters like cuisine or star rating, the system is designed to understand more nuanced inputs such as:

“somewhere cozy but not too expensive”
“spicy food, but not too oily”
“good for a group with mixed tastes”
“authentic local food near us”

We also connected restaurant and location data through mapping / restaurant APIs so recommendations could be grounded in real places, not just generated ideas. The result is a system that feels more conversational, more contextual, and more useful for actual decision-making.

AI plays a central role in the Table-Us experience. It interprets vague natural-language inputs, translates them into structured taste and context signals, balances individual preferences with group compatibility, and helps narrow large sets of restaurant options into more meaningful recommendations. Instead of using AI as a cosmetic add-on, we designed it as the core decision layer that makes social dining faster, smarter, and more personalized.

Challenges we ran into

One of our biggest challenges was defining what “taste” really means in a product. Taste is not just cuisine type. It includes texture, spice level, atmosphere, familiarity, dietary constraints, social setting, and even emotional intent. Translating that into a usable interface and an AI-friendly structure was much harder than using standard restaurant filters.

Another challenge was balancing individual preference with group compatibility. A great solo recommendation is not always a great group choice. We had to think carefully about how an AI system should mediate tradeoffs instead of just optimizing for one person.

We also spent a lot of time on product framing. We did not want Table-Us to feel like “just another restaurant app with AI added on top.” We wanted the AI to be central to the value: helping users communicate preferences better, discover more meaningful options, and make decisions together with less friction.

What we learned

We learned that food discovery is not only a search problem; it is also a coordination problem, a culture problem, and a personalization problem.

We also learned that AI works best here when it does not replace choice, but clarifies it. The most valuable role for AI is not to overwhelm users with more options, but to reduce ambiguity and help groups move toward confident decisions.

Finally, we learned how important product storytelling is. A strong technical implementation matters, but for a project like this, the user pain has to be immediately understandable. The best ideas are the ones people instantly recognize from their own lives.

What's next for Table-Us

Next, we want to make Table-Us even more social and more context-aware.

Our future plans include:

richer taste profiles that evolve over time
smarter group matching and compromise suggestions
stronger local and cultural discovery features
voice-based preference input
better integrations for reservations, invites, and planning

Our long-term vision is to make Table-Us the social layer for dining decisions: a product that helps people not just find food, but find the right food, with the right people, in the right moment.

How we used Cursor

Throughout the project, we used Cursor as a core development tool to accelerate both engineering and product iteration. Cursor helped us move faster when building UI components, refining frontend flows, debugging integration issues, and restructuring code as the product direction evolved.

We also used Cursor to support rapid prototyping and AI-assisted development. It was especially useful for generating implementation scaffolding, refactoring repeated code patterns, improving code readability, and helping us iterate on prompts and system behavior more quickly. Instead of treating Cursor like simple autocomplete, we used it as an active coding partner to shorten the path from idea to working feature.

Most importantly, Cursor helped us keep momentum during the hackathon. Because Table-Us combines interface design, AI reasoning, restaurant discovery, and group decision logic, we needed to iterate across multiple layers of the product very quickly. Cursor made it easier to test ideas, revise flows, and ship a more polished prototype in less time.